Because typhoons are among the most devastating natural phenomena on earth, the meteorologists who monitor and predict their movements occupy a position of particular respect. The experts at the ROC Central Weather Bureau who specialize in this task are called the "Typhoon Team.”
Before the team was established, mid-April to mid-November was an annual period of worry and even fear among residents, especially in the southern part of the island where typhoons are more frequent and severe.
Today, thanks to the Typhoon Team, residents can rely on high technology and professional expertise to help protect them from being caught unawares and unprepared for the high wind, heavy flooding, and serious landslides common during the storms.
Sarcastic remarks about the accuracy of Central Weather Bureau predictions are as common in Taipei as they are in New York. Weekend plans made on the basis of daily weather forecasts can be frustrated by sudden changes in temperature unexpected by weathermen as much as by those planning a drive in the countryside. Because meteorology remains an inexact science, the ROC Typhoon Team can be especially proud of its record, which is one of the best in the world.
The team is headed by Director General Wu Tsung-ya of the Central Weather Bureau (CWB), along with Director Shieh Shinn-liang of the CWB's Forecast Center. Shieh supervises the staff, which consists of a deputy director, several section chiefs and senior forecasters, and a draftsman who plots weather charts.
Shieh and Forecast Section Chief Kenneth Chen also work regularly with a large group of non-team meteorologists, computer and communications specialists, and other technicians to determine day-to-day weather patterns. One of their major preoccupations is finding clues that indicate a new typhoon is in the making. Although storms have hit the island as early as mid-April, the most dangerous months during the typhoon season are July, August, and September.
Typhoons are moody entities, and they often thwart the predictions of top experts who have a thorough knowledge of subtropical air masses and general weather patterns. A central problem concerns timely specific information. "One of our most serious drawbacks in plotting the behavior of a typhoon is a lack of data on local weather conditions far out at sea," Shieh explains. "But we can keep an eye out for typical typhoon conditions. These storms usually start in large areas of low-level vortex (wind circulation) over the North Pacific Ocean, where the surface water temperature exceeds 26 degrees Celsius and the temperature in surrounding areas is lower. "
The higher temperature can push down barometric pressure and form what meteorologists call a low-pressure trough. The difference in air pressure may cause clouds to start whirling around a center, or eye. The whirling in turn forces the pressure even lower. The result is a typhoon, called tai feng in Chinese, which means great wind.
Taiwan is just north of the prime source of typhoons in the northwestern Pacific Ocean. This location, combined with the prevailing westerly winds slightly to the north, means that every summer and fall the island lies directly in the path of about half a dozen of the region's storms. The east coast and the southern area of Taiwan are prime targets for the three or four typhoons that normally reach the interior each year.
When local weathermen spot one of the whirling depressions formed by a low-pressure trough, Shieh puts his team on standby. "Our first priority is to determine if it might affect this area," he says. "The team meets at least 48 hours before we expect the first sea warning will be issued, which means 24 to 36 hours before the periphery of the storm approaches within 100 kilometers of the island. So we've had our eyes on the typhoon for two or three days before the public hears about it."
Three days of watching means three days of camping-out in the office. The team members take turns sleeping, sometimes at their desks. "The longest we've ever had to be here without going home was 15 days," Director General Wu says. "That was three years ago in August, when Typhoon Wayne hit Taiwan then swept back across the island twice before dissipating. It was the worst typhoon in the island's history. Most only last about four or five days."
1988 was strangely quiet on the storm front. The summer was unseasonably cool, with prolonged rains and severe flooding, but only one typhoon directly hit the island. Farmers enjoyed a rare season free from storm-damaged crops, while city-dwellers avoided weeks of flooded streets, power outages, and high-priced vegetables. But experts were perplexed by the season, and were uncertain about the causes or the influences on weather patterns this year.
During the summer, a subtropical high pressure air mass usually hovers over Taiwan and the rest of southeast Asia. But last year it remained much farther south. And the first cold front to descend from the north came in July, two months earlier than normal. This was the earliest recorded arrival of the cold air mass in Taiwan. Adding to the anomalous situation, the Indian monsoon that drenches much of Asia during the summer last year pushed eastward all the way to Okinawa, subjecting Taiwan to unseasonal rains.
"The unusual location of the sub-tropical high pressure air mass, combined with the effects of the Indian monsoon front, were probably the reasons for last year's phenomenon," Chen says. "Typhoon formation is typically along the subtropical high southern edge, also called the intertropical convergence zone (ITCZ). At the ITCZ, high clockwise-moving subtropical winds in the northern hemisphere collide head-on with air masses moving counter-clockwise in the southern hemisphere."
These winds are a result of the rotation of the earth itself. "This confrontation, combined with the high moisture and temperature typical of the tropics, gives rise to violently whirling air masses, which can accelerate until they become typhoons," Chen continues. "But the ITCZ was too far south in 1988. Typhoons therefore traveled up along the northern edge of the subtropical high, south of Japan, at higher latitudes that are much drier and colder. Under such conditions, typhoons are weaker."
The paths of last year's storms were also unusual. "Instead of heading north west, the usual direction for typhoons in this area, these high-latitude storms went north, or even northeast, away from Taiwan," Chen says. "That direction was generated perhaps by the storms meeting the cold front, or low-pressure zone, and the monsoon coming from the west."
The only threatening storms last year were Typhoon Susan in early June and Typhoon Warren in mid-July. Although they demanded marathon watch sessions of three days each and forced the Weather Bureau to issue warnings, only the latter hit the island.
The Typhoon Team's main job is to predict the paths of storms. But the challenging task of forecasting is made even more difficult by inadequate information concerning weather patterns over the ocean. Most data is derived from weather balloons, observation stations on small islands [see FCR, November 1987], and satellite pictures. Large ships are also required to radio information every three hours on temperature, pressure, wind speed and direction, precipitation, cloud cover, and other conditions.
"Satellite photos give detailed information on the structure of a storm," Shieh says. "High-altitude reconnaissance aircraft from the U.S. Air Force base on Guam also supply valuable data. Pilots risk their lives by flying directly through the typhoons to collect information unavailable on the periphery. Most important of all, they locate the eye of the storm, which is usually 30 to 50 kilometers in diameter. The radius of a typical typhoon is 200 to 300 kilometers with an altitude of 10 to 16 kilometers."
As a typhoon approaches Taiwan, radar scopes provide continual data. Over the last few years, radar equipment at stations in Kaohsiung in the south and Hualien in the east have been upgraded for more accurate tracking. The radar system is still incomplete, however, because northern Taiwan still lacks a station, pending the choice of an appropriate site. The Weather Bureau is radar-blind if a storm takes a northern route.
Processing the mountains of data compiled from all these sources requires a high-speed, large-memory computer. Team members use data from past storms to find clues to the behavior of current and future typhoons. Statistical predictions draw on information about the direction, speed, location, and timing of storms dating back to 1959.
But one adviser to the Forecast Center, Dr. Tsay Ching-yen, who was formerly chairman of the Department of Atmospheric Science at National Taiwan University, points out that such predictions are only good for the average typhoon. "Many typhoons have a will of their own," he says. "Statistical predictions can be made more accurately by using concurrent wide-range meteorological data, but that too is of limited value since conditions continually change."
Another promising approach to storm prediction involves the use of complex, physics-based equations, such as those governing the movement of fluids. These equations can provide dynamic models of weather patterns superior to single-moment statistics or current data. "The dynamic model allows you to predict the future of our atmosphere," Tsay explains. "But these types of models are still experimental. Most are still being tested by weather bureaus around the world."
The Typhoon Team is in fact using a dynamic model Tsay himself developed and introduced 10 years ago. The model is based upon another equation, but was modified to take into account weather conditions peculiar to Taiwan. It was first applied to an actual situation three years ago, after the team obtained the computer equipment necessary for the task.
"The complexities of this mathematical model demand a long trial phase," Chen says. "We still must learn more about how it actually works and modify it accordingly. But it has already proved helpful, although the predictions tend to run a bit late."
Paradoxically, the statistical and dynamic methods of prediction often yield conflicting results. A diagrammatic study conducted two years ago to predict the path of Typhoon Alex used both methods, and they yielded a host of contradictory directions. Four equations indicated the storm would curve east and bypass Taiwan, while two suggested that it would hit the southern part of the island. "Just a one-degree difference in the calculations can sharply affect the predictions, giving opposite results," Chen explains. Shieh adds that individual typhoons have their own idiosyncrasies, and a degree of experience-based guesswork is unavoidable. "This is the subjective method, and I have used it many times over the last 20 years," he says.
Whenever a typhoon is forming in the Pacific, team members meet every three hours to compare notes and make predictions. Since most of the weathermen have been on the job more than 10 years, they have an experienced sense of weather patterns. If a typhoon approaches within 400 kilometers of Taiwan, the team meets every hour, which is the minimum time required to plot a path. For unusual storms, a forecast may in fact require several hours to formulate.
Shieh and Chen enjoy a good record, and believe it will improve. When Typhoon Susan was in the vicinity last June, the team produced far better forecasts than either the weather bureau in Japan or the U.S. Joint Typhoon Warning Center on Guam. The Japanese and Americans predicted Susan would proceed directly to mainland China once west of Taiwan, but the Typhoon Team predicted Susan would backtrack to the island. "The team proved correct," Shieh says. "As storms enter our area we now make predictions with greater confidence. "
By the end of this year, the CWB's US$4.9 million super-computer will be completely upgraded, and will occupy more than two floors. Already programmed for global forecasting, it will be the most powerful system in the country. Work is already under way to program it for regional and meso-scale weather patterns. But the crowning achievement will be its "typhoon track model" program.
Shieh hopes that by the time the computer is fully operational his team will have fully implemented both Tsay's original mathematical model and a few others developed by himself and his staff. With such impressive back-up technology, the day is nearing when ROC weathermen may upgrade weather prediction into much more than an inexact science.